Appendix 4: Interventions Aimed at Professionals

Appendix 4: Interventions aimed at professionals

SR 1st Author (Year) / Title of SR / Study designs in review (#) / Intervention details / Results
Educational Outreach Visits & Audit and Feedback Interventions, n = 4
Goldberg (2007) [36] / Pain Management in Hospitalized Cancer Patients: A Systematic Review / RCT (1)
pre-post (3) / •improve management of pain
•cancer patients only
•participants - nurses
•various educational outreach tactics / •increased patient satisfaction
•increased documentation
•increased nursing knowledge
•no improvementwith patients’ pain severity ratings
O’Brien (2008) [44] / Educational outreach visits and effects on professional practice and health outcomes / RCT (69) / •improve clinical practice (varied practice targeted)
•mixed population with cancer
•participants - physicians, nurses, other providers
•various educational outreach tactics / Outcomes / EOA part vs None / EOA alone vs None
provider practice dichotomous outcomes
ARD median improvement / 5.6%: -3% to 64% / 5.0%: 1% to 20%
ARD < 5% / 15 of 34 / 9 of 21
ARD 5%-9% / 11 of 34 / 8 of 21
ARD 10+% / 8 of 34 / 1 of 21
provider practice continuous outcomes
A % median improvement / 21%: 0% to 61.4% / 23%: 0% to 617%
A% <10% / 4 of 20 / -
A% 11-20% / 5 of 20 / -
A% 20%+ / 9 of 20 / -
patient outcomes / few studies, small effects to no effects
•EOV at in individual level tend to be more effective than in groups
•multifaceted interventionsof which EOV is a part tends to be more effective than single interventions
•interventions that include EOV tend to be more effective than audit and feedback alone
•EOV yields small but consistent benefits for prescribingbehavior in contrast to other behaviors where effect sizes vary considerably
Goldberg (2007) [36] / Pain Management in Hospitalized Cancer Patients: A Systematic Review / pre-post (3)
CCT (1) / •improve management of pain
•cancer patients only
•participants - nurses
•audit and feedback / •improvements on assessment rates, provider satisfaction but no impact on clinical outcomes (patient pain scores)
Jamtvedt (2006) [40] / Audit and feedback: effects on professional practice and health care outcomes / RCT (118) / •improve clinical practice
•mixed population with cancer
•participants - physicians, nurses, other providers, units, departments, institutions
•audit and feedback using various tactics / Comparison / Provider Practice
dichotomous
ARD improvement / Provider Practice
continuous
A% improvement
AF Part vs. None / 5.0%: -16% to 70% / 5.0%: -10% to 68%
AF Alone vs. None / 4.0%: -16% to 32% / 11.9%: -10.3% to 67.5%
AF + EOV vs. None / 1.5%: -1% to 24% / 28.7%: 3% to 41%
AF as part of multifaceted (not EOV) vs. None / 5.7%: -9% to 70% / 23.8%: 3% to 60%
•very few factors could explain variability of effects
•AF may be more effective when baseline compliance is poor though the effect is not large
•higher intensity of AF is more effective than lower intensity AF
•there is little data to direct how AF should be delivered
•no evidence to support or refute AF more effective with the addition of peer performance data
•no evidence to support or refute AF more effective when provider motivation is high (vs. low or medium)
•AF may be more effective when outcomes are considered serious
IT/IM/Informatic Interventions, n = 5
Goldberg (2007) [36] / Pain Management in Hospitalized Cancer Patients: A Systematic Review / CCT (1) / •improve management of pain
•cancer patients only
•participants - physicians
•CDSS that provide recommendations to physicians regarding drug options, dose escalation, modifying agents, adverse effects, etc. / •appropriate reduction in prescribing a particular drug
•no change in pain scores
Garg (2005) [34] / Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review / RCT (49) CRCT (38) TS (1)
other (11) / •improve clinical care and patient outcomes across clinical conditions
•mixed population with cancer
•participants – physicians, interns, residents
•interactive CDSS whereby clinical staff or patients enter patient characteristics and receive advice/recommendations / •clinical performance improved in 64% of 97 studies
•40% using systems targeting diagnosis
•76% using reminder systems
•62% using disease management systems
•66% with using drug dosing and prescribing systems
•patient outcomes improved in 13% of 52 studies
•general findings
•automatic prompt systems greater impactthan user activated prompt systems more effective (73% vs. 47% of trials)
•systems developed by authors more effectivethan systems in which authors were not the developers (74% vs. 28% of trials)
Ammenwerth (2008) [23] / The effect of electronic prescribing on medication errors and adverse drug events: a systematic review / RCT (1) CRCT (1)
TS (6)
PC (6)
RC (2)
CS (1)
pre-post (9) / •to improve patient safety across clinical conditions
•mixed population with cancer
•participants – physicians and patients
•no decision support (provides a list of drugs, doses, monographs) vs. limited decision support (evidence-based specific recommendations of drugs, doses, etc.) vs. advanced decision support (drug allergy or interaction information, drug-lab alerts, etc.) / •medication error rate
•reduction in error rate for 23/25 studies (RR range 13% to 99%)
•inconclusive findings for 1/25 studies
•increase risk in error rate for 1/25 studies
•adverse drug events (potential)
•reduction in adverse drug events for 6/9 studies (RR range 35% to 98%)
•inconclusive findings for 2 of 9 studies
•increase in adverse drug event for 1/9 studies
•adverse drug events (actual)
•reduction in adverse drug events for 4/7 studies (RR range 30% to 84%)
•inconclusive finding for 1 of 7 studies
•increase in adverse drug event for 1/7 studies (9%)
Beach (2006) [25] / Improving health care quality for racial/ethnic minorities: a systematic review of the best evidence regarding provider and organization interventions / RCT (7)
CCT (3) / •improve quality of health care and reduce disparities
•mixed population with cancer
•participants – physicians, residents, fellows, nurses
•cancer mix / •all studies showed improvements in appropriateness of care (e.g. prevention of preventive care, tobacco cessation counseling, etc.)
Shojania (2010) [50] / The effects of on-screen, point of care computer reminders on processes and outcomes of care (review) / RCT (2)
CRCT (20)
CCT (4)
CCCT (2) / •mixed population with cancer
•participants – physicians, nurses, pharmacists, dentist
•computer system reminders alerting providers to potentially redundant tests / •consistent but small median absolute improvements in process adherence
•across all outcomes 4.2% (IQR 0.8% to 18.8%)
•ordering of medication, 3.3% (IQR 0.5% to 10.6%)
•prescription of vaccines 3.8% (IQR 0.5% to 6.6%)
•test ordering 3.8% (IQR 0.4% to 16.30%)
•documentation 0% (IQR -1.0% to 1.3%)
•other 1% (IQR 0.8% to 8.5%)
•larger effects if best outcome is chosen
•median improvementin dichotomous clinical endpoints 2.5% (1.3% to 4.2%)
•general
•effect sizes larger with quasi-randomized (vs. randomized), in-patient (vs. outpatient), reminders alone (vs. part of multifaceted)
•no differencein effect size regardless of baseline compliance or specific tactics of reminder system
Local Opinion Leader Interventions, n = 1
Doumit (2006) [31] / Local opinion leaders: effects on professional practice and health care outcomes / RCT (12) / •improve clinical practice across clinical conditions
•mixed population with cancer
•participants – physicians, nurses, midwives
•local opinion leaders responsible for leading educational initiatives / •decrease in rates of non-compliance:
•LOL vs. None: median ARD 7% (-6% to 12%)
•LOL vs. Single Intervention: median ARD 14% (12% to 17%)
•LOL + Additional Intervention vs. AI only: median ARD 9% (2% to 25%)
•LOL as part of multifaceted approach vs. None: median ARD 6% (1% to 14%)
•general
•insufficient data to inform most effective method of choosing LOL or educational tactic
•poorer quality studies yielded bigger effect sizes than higher quality studies
Tailored Interventions, n =1
Baker (2010) [24] / Tailored interventions to overcome identified barriers to change: effects on professional practice and health care outcomes / RCT (26) / •improve treatment decisions across and clinical outcomes across clinical conditions
•no cancer
•participants - physicians, nurses, pharmacists, and health teams
•tailored interventions designed to mitigate identified barriers / •pooled OR 1.54 (95% CI, 1.16 to 2.01) in favour of tailoring (binary clinical outcomes)
•summary
•8/14 studies demonstrated benefit of tailoring
•2/14 studies demonstrated some benefit of tailoring for some outcomes
•4 of 14 studies showed no benefit of tailoring
•general
•unclear how tailoring was done in many cases (or if it was done)
•little data if tailoring overcame barriers it aimed to mitigate
•no study attributes (quality attributes, level of tailoring, complexity of behavior) were associated with interventions’ effectiveness
Clinical Pathways, n =1
Rotter (2010) [46] / Clinical pathways: Effects on professional practice, patient outcomes, length of stay and hospital costs / Various comparison studies (27) / •impact on professional practice, patient outcomes, length of stay, and costs
•mixed population including cancer
•participants - physicians, nurses, allied health providers
•varied formats, modalities, adherence to defining elements / •CP alone vs. usual care
•in-hospital complications: reduced (OR 0.58; 95% CI 0.36-0.94)
•documentation: improved (OR 13.65; 95% CI 5.38 to 36.64)
•rates of hospital re-admission: n.s.
•rates of in-hospital mortality: n.s.
•length of stay: most studies reported reduction
•hospital costs/charges: WMD = $-4919 US
•CP as part of multifaceted intervention vs. usual care
•no significant differences on outcomes
Guidelines in Professions Allied to Medicine, n = 1
Thomas (2009) [53] / Guidelines in professions allied to medicine. / Various comparison studies (18) / •impact of guidelines on processes of care, outcomes of care, and costs
•mixed population including cancer
•participants - nurses, allied health providers
•varied formats, modalities, adherence to defining elements / •guidelines + dissemination/implementation strategies vs. no guidelines with dissemination or implementation strategy
•processes of care: improvements with at least one indicator in 3 of 5 studies with outcome
•outcomes of care: improvements with at least one indicator in 6 of 8 studies
•insufficient evidence to advise on specific guideline dissemination or implementation tactics
•insufficient evidence to advise on financial impact to health care system
Discharge Planning from Hospital to Home, n = 1
Shepperd (2010) [49] / Discharge planning from hospital to home / RCT (21) / •impact of DP on length of stay, readmissions, patient outcomes, satisfaction, costs
•mixed population including cancer
•participants - patients, nurses, pharmacists, physicians
•varied formats, modalities, adherence to defining elements but most had assessment, implementation of plan, and monitoring
•impact on mortality, health outcomes and cost are less certain / •hospital length of stay: small reduction (MD -0.91, CI 95% -1.55 to -0.27)
•readmission: small reduction (RR 0.85: CI 95% 0.74 to 0.97)
•satisfaction: 3 studies report increased patient satisfaction
•impact on mortality, health outcomes and cost are less certain

NB: SR = systematic review, RCT = randomized controlled trial, DS = descriptive, non-RCT = non-randomized controlled trial, RT = non-controlled randomized trial, TA = tandem assignment, LR = literature review, MM = mixed methods, 2x2 FC = 2x2 factorial comparison, NS = not specified, CRCT = cluster RCT, CCT = controlled clinical trial, TS = time series, PC = prospective cohort, RC = retrospective cohort, CS = cross-sectional, CCCT = cluster controlled clinical trial, QEL = quasi-experimental using linear modeling, CE = cost-effectiveness, RS = relationships, CBA = controlled before-after, OBS = observational, CC = case control